Part of #5821 .

Description

The main idea of this PR is to solve 3 pain points of the current bundle training or inference workflow:

1. Training or inference is defined in a "rigid" training or evaluation list, so 3rd party package can't flexibly initialize and run the workflows separately, see: Define initialize, run, finalize sections in bundle config model-zoo#294. This PR splits it into initialize, run and finalize, and still compatible with current existing bundles.
2. 3rd party packages need to access the bundle config directly with specifed keys based on ConfigParser, but some developers of bundles don't want to write JSON or YAML configs. This PR implemented the interface to support both config-based and python-based workflows.
3. MONAI Label, MONAI deploy, MONAI-FL require many fixed properties of a training or inference workflow, but we don't have definition for them in the code, just described in the developer guide doc of model-zoo. This PR implemented the TrainProperties and inferenceProperties interfaces and check tool.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.